Dear Lee,
Many thanks for your help, I think I solved the problem with your suggestion
to use "egen, concat". Vars m1-m6 are coded 0/1 so I started with the
following:
for X in num 2/6: recode mX 1=X
egen new=concat(m1 m2 m3 m4 m5 m6)
and I did get the desired output as follows:
new | Freq. Percent Cum.
-----------------------------------------------
000000 | 21 7.07 7.07
100000 | 2 0.67 7.74
103000 | 1 0.34 8.08
103456 | 3 1.01 9.09
123000 | 3 1.01 10.10
123400 | 2 0.67 10.77
123450 | 2 0.67 11.45
123456 | 263 88.55 100.00
------------+-----------------------------------
Total | 297 100.00
Once again, my thanks for your good suggestion and time.
Amani
-----Original Message-----
From: Lee Sieswerda [mailto:[email protected]]
Sent: Saturday, 23 August 2003 00:00
To: '[email protected]'
Subject: st: RE: Using egen or similar
I think this will give you pretty close to what you want.
* Reduce your master data to only the six
* variables of interest, and concatenate
* into a new variable
use your_master_data.dta
keep m1-m6
egen bincats = concat(m1-m6)
keep bincats
sort bincats
save temp1.dta, replace
* Generate a new dataset with all
* possible categories, in this case
* you've got 2^6 possible categories
clear
set obs 64
gen cat1 = 0
gen cat2 = 0
gen cat3 = 0
* Okay, this block is a hack job. I know
* there is a nice algorithm for
* generating all of the permutations
* of a binary outcome over n events, but
* I can't put my hands on it right now
replace cat1 = 1 in 33/64
replace cat2 = 1 in 17/32
replace cat2 = 1 in 49/64
replace cat3 = 1 in 9/16
replace cat3 = 1 in 25/32
replace cat3 = 1 in 41/48
replace cat3 = 1 in 57/64
egen cat4 = fill(0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1)
egen cat5 = fill(0 0 1 1 0 0 1 1)
egen cat6 = fill(0 1 0 1)
egen bincats = concat(cat1-cat6)
* Now generate the kind of
* labels you want, more or less
forvalues n = 1/6 {
replace cat`n' = cat`n'*`n'
}
egen numcats = concat(cat1-cat6)
keep numcats bincats
* Now merge your original data back into these
* categories
sort bincats
merge bincats using temp1.dta
* Drop the categories not needed
drop if _merge==1
* And tabulate
tab numcats
There may also be a clever way to do this with string functions rather than
with -merge-.
Regards,
Lee
Lee Sieswerda, Epidemiologist
Thunder Bay District Health Unit
[email protected]
-----Original Message-----
From: [email protected] [mailto:[email protected]]
Sent: Friday, August 22, 2003 2:43 PM
To: [email protected]
Subject: st: Using egen or similar
Dear Stata-listers,
I have six binary variables (m1,..,m6) coded 0/1
I wish to see the frequency of the pattern of occurring 1s, but preserving
the variable sequence. I am not sure whether any of "egen" functions can
perform the task (and I prefer not to use egen..=group(..)).
I started off with the following:
for X in num 2/6: recode mX 1=X
then if there is one of egen functions it would be
egen new=fun(m1 m2 m3 m4 m5 m6)
then tabulate new would yield
new freq
------ ------
1,2,3 63
1,3,4,5 23
1,4,5 30
2,5,6 20
.....etc
It is if I am "xtdes" but with variables rather than panel time-entries.
Many thanks for your help in advance.
Amani
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
